97P - Neoadjuvant durvalumab plus gemcitabine and cisplatin (D+GemCis) versus gemcis alone for localized biliary tract cancer (BTC): Results of a randomized, multicenter, open-label, phase II trial (DEBATE)

Background

D+GemCis demonstrated significantly improved efficacy in patients (pts) with advanced/metastatic BTC compared to GemCis alone. However, the efficacy of D+GemCis in the neoadjuvant setting has not been studied.

Methods

This was a multicenter, open-label, non-comparative, randomized phase 2 trial including treatment-naïve pts with histologically or cytologically confirmed localized BTC. Pts were randomly assigned (2:1) to receive either 4 cycles of neoadjuvant D+GemCis or GemCis alone. Postoperative 6 cycles of D were given if pts underwent surgery regardless of assigned neoadjuvant therapy. The primary endpoint was R0 resection rate. ctDNA analysis was performed using GuardantOmni platform.

Results

A total of 45 pts were enrolled (n=31 in the D+GemCis arm and 14 in the GemCis arm). Intrahepatic cholangiocarcinoma (n=18, 40%) was the most common primary tumor site and 69% (n=31) had clinical stage III. Response rates were 36% (n=11) in the D+GemCis arm and 7% (n=1) in the GemCis alone arm. In the D+GemCis and GemCis arms, surgical exploration was performed in 68% (n=21) vs 36% (n=5), curative-intent resection (R0+R1) in 61% (n=19) vs 36% (n=5), and R0 resection in 48% (n=15) vs 36% (n=5), respectively. Median PFS were 15.1 mo (95% CI, 4.5-25.8) in the D+GemCis arm and 3.6 mo (0-13.0) in the GemCis alone arm. Pts able to receive curative-intent surgery after study treatment had significantly better overall survival (p<0.001). No new safety signal for D+GemCis were identified. ctDNA analysis revealed that 5 pts (11%) had actionable genetic alterations including 4 with IDH1 R132C and 1 with ERBB2 amplification, and 2 pts (4%) were MSI-H.

Conclusions

Neoadjuvant D+GemCis resulted in a higher surgical resection rate in pts with localized BTC, and surgical resection was associated with better survival. ctDNA analysis was shown to be a feasible method for assessing actionable target gene alterations in early BTC.

Clinical trial identification

NCT04308174.

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

AstraZeneca and CKD Pharmaceuticals.

Disclosure

C. Yoo: Financial Interests, Personal, Invited Speaker: Bayer, Celgene, Eisai, Ipsen, Servier, Roche, Novartis; Financial Interests, Personal, Advisory Board: AstraZeneca; Financial Interests, Personal, Research Grant: AstraZeneca, Bayer, Servier; Financial Interests, Institutional, Research Grant: Genentech. J.O. Park: Financial Interests, Personal, Advisory Board: MedPacto, BMS (Celgene), Servier, MediRama, Adicet Bio, AstraZeneca, Merck Sereno; Financial Interests, Personal, Other, Travel support for a poster presentation at ASCO GI 2023: Minneamrita Therapeutics LLC; Financial Interests, Personal, Research Grant, Clinical research grant: MedPacto, Servier, BMS (Celgene); Financial Interests, Personal, Research Grant: Eutilex, ABL Bio. All other authors have declared no conflicts of interest.